Steam API rate limits, and how to stay under them.

Valve documents three things: a daily limit of 100,000 calls per key, the 429 Too Many Requests response for rate-limited callers, and stricter per-IP limits for traffic that keeps producing 403 errors. Anything beyond that, such as a per-second ceiling, is not published.

100,000 a day averages to about 69 calls a minute, or 1.16 a second, spread evenly. Real traffic is bursty, so plan for a sustained rate well below that and leave headroom for retries.

Cache each response for as long as the data stays valid, batch IDs where the method accepts a list (GetPlayerSummaries takes up to 100 Steam IDs), refresh only what users actually look at, and retry with exponential backoff instead of immediately.

import time, threading

class DailyBudget:
    """Spread a daily allowance evenly; block when it is used up."""
    def __init__(self, per_day=90_000):
        self.rate = per_day / 86_400          # tokens per second
        self.capacity = 60                    # allow short bursts of 60 calls
        self.tokens, self.t = self.capacity, time.monotonic()
        self.lock = threading.Lock()

    def take(self):
        with self.lock:
            while True:
                now = time.monotonic()
                self.tokens = min(self.capacity, self.tokens + (now - self.t) * self.rate)
                self.t = now
                if self.tokens >= 1:
                    self.tokens -= 1
                    return
                time.sleep((1 - self.tokens) / self.rate)

budget = DailyBudget()
# call budget.take() before every request to api.steampowered.com

The CS2 API publishes per-minute and monthly limits for every plan, and every 429 carries Retry-After, X-RateLimit-Limit, X-RateLimit-Remaining and X-RateLimit-Reset, so a client always knows when to try again. Pro esports data is not in Steam at all, so for match data the comparison is moot.